Legal claims defining the scope of protection, as filed with the USPTO.
1. An audio signal coding apparatus which receives an audio signal which is subjected to time-frequency transform to output index information, comprising: a plurality of stages of sub-quantization units, at least the i-th stage sub-quantization unit including a plurality of band division quantizers each of which being operable to perform coding using vector quantization for a respective part of a divided frequency band; each of the (i+1)th and subsequent stage sub-quantization units being operable to perform coding using vector quantization on a quantization error of the band division quantizer that has performed decoding on a part of the band at the previous stage; and a quantization band selection unit operable to select the parts of the divided frequency band for quantization by the band division quantizers in the (i+1)th and subsequent stage sub-quantization units, on the basis of the magnitude of the quantization errors from a previous stage sub-quantization unit; wherein: the index information includes a stage-specific index indicating a vector coded by the sub-quantization unit of each stage; and the stage-specific index includes a band-specific index for each part of the divided frequency band at the stage in which codings with dividing the frequency band are performed by the plurality of band division quantizers.
2. An audio signal decoding apparatus for decoding index information into an audio signal, comprising a plurality of stages of sub inverse quantization units, at least the i-th stage sub inverse quantization unit including a plurality of band division inverse quantizers each of which being operable to perform decoding using vector inverse quantization for a respective part of a divided frequency band; each of the (i+1)th and subsequent stage sub inverse quantization units being operable to perform decoding using vector inverse quantization on a quantization error of the band division inverse quantizer that has performed decoding on a part of the band at the previous stage; wherein: the index information includes a stage-specific index indicating a vector for each stage; the stage-specific index includes a band-specific index for each part of the divided frequency band at the stage in which decoding with dividing the frequency band is performed; and after decoding by a first band division inverse quantizer in the i-th stage sub inverse quantization unit is performed, and before decoding by a second band division inverse quantizer in the i-th stage sub inverse quantization unit is performed, decoding by the (i+1)th stage sub inverse quantization unit is performed.
3. An audio signal decoding method for decoding index information into an audio signal, comprising: a plurality of stages of sub inverse quantization steps, at least the i-th stage sub inverse quantization step including a plurality of band division inverse quantization steps to perform decoding using vector inverse quantization for a respective part of a divided frequency band; in the (i+1)th and subsequent stage sub inverse quantization steps, decoding using vector inverse quantization is performed on a quantization error generated in the band division inverse quantization step that has performed decoding on a part of the band at the previous stage; wherein: the index information includes a stage-snecific index indicating a vector at each stage; the stage-specific index includes a band-specific index for each nart of the divided frequency band at the stage in which decoding with dividing the frequency band is performed; and after decoding by a first band division inverse quantization step in the i-th stage sub inverse quantization step is performed, and before decoding by a second band division inverse quantization step in the i-th stage sub inverse quantization step is performed, decoding by the (i+1)th stage sub inverse quantization step is performed.
Unknown
July 10, 2007
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.